Written by: Techub News Compilation

In a recent interview with Y Combinator, serial entrepreneur and Sasabi founder Shomik shared his journey from the medical AI sector to building an AI-native observability platform. As a member of the early infrastructure and observability team at Brex, and founder of the startup Opkit, Shomik combined his deep technical background with entrepreneurial lessons to present a clear vision for next-generation developer tools: to completely eliminate the cumbersome manual debugging processes with AI, moving software towards self-repair. This conversation not only revealed how AI is reshaping infrastructure tools but also provided valuable reflections for tech entrepreneurs.

From Opkit to Sasabi: A Return to the Core of Entrepreneurship

Shomik's first startup, Opkit, was an AI voice assistant company focused on the healthcare sector, aimed at automating phone communications with insurance companies, such as eligibility verification, pre-authorization, and claims status inquiries. This idea stemmed from his background in a family with a doctor and his judgment that "vertical fintech" could become the next big wave. However, reflecting on this experience, Shomik admitted that Opkit felt more like an "MBA case study" entrepreneurial choice—based on rational analysis of market opportunities rather than personal passion, experience, or deep insights.

Despite the team's technical progress, even becoming one of the first teams to commercialize LLM voice assistants in healthcare, the entire entrepreneurial journey was fraught with challenges. Shomik reflected that they were not the “right team” to start a medical AI company, although they later accumulated a wealth of domain knowledge through hard work. He recognized a key lesson: entrepreneurship should ideally be based on the founders’ own expertise, experience, and passion rather than deliberately avoiding familiar areas in pursuit of seemingly grand but unfamiliar opportunities. One of his biggest regrets about Opkit was not examining earlier and more deeply whether "this truly played to our strengths" and "whether we were willing to fully commit to the healthcare technology industry for the next decade."

After operating Opkit for about two years, the team faced financing challenges and decided to seek acquisition. Ultimately, they joined a rapidly growing AI sales technology company, 11x. It was his experience at 11x that served as a direct catalyst for Shomik to establish Sasabi. As the tech lead, he was responsible for rebuilding the company’s AI sales development representative product. After the project reached a stable maintenance phase, he found himself once again caught in a familiar dilemma: the team was using cutting-edge AI coding tools to develop futuristic products, but when production issues arose, the debugging process was still the painful, inefficient manual mode of the last decade. This stark contrast made him realize that in the realm of software maintenance and observability, a transformation driven by AI was overdue but essential.

“I realized that this is the company I should be founding,” Shomik stated. Sasabi completely aligns with his personal qualities: it targets developers like him, addressing pain points he has personally experienced (observability), and its customers are early-stage and growth-stage tech startups (the only type of company he has ever worked for), even the company name derives from his favorite sci-fi anime "Mobile Suit Gundam." He sees this as a symbolic action to ensure everything in the new company aligns with “who I am.” Compared to the arduous journey during Opkit, he is determined to make the entrepreneurship process at Sasabi "fun."

"Logs Are Everything": How AI Reshapes the Philosophy of Observability

Sasabi positions itself as an "AI-native observability platform," designed specifically for rapidly iterating engineering teams. Shomik likens it to "a DataDog or Sentry built in 2026," with the core feature being deeply integrated AI agents. The platform allows developers to ask questions directly in natural language, such as "Why is the production environment down?" "What does this error mean?" "Which customers are affected?" or "Which commit caused it?" thereby quickly pinpointing and resolving production issues.

The most controversial yet central concept is Sasabi's proposition of "Logs Are Everything." This directly challenges the long-held "three pillars" theory (logs, metrics, traces) in the observability field. The traditional view posits that these three types of telemetry data serve different querying needs, each with its volume and efficiency characteristics, thus requiring engineering teams to implement all three concurrently, leading to significant complexity and burden.

Shomik offers a different perspective:

• Logs have an extremely low adoption threshold: Every developer knows how to write print statements and read log streams, making it the simplest and most direct way of observability, practically the "Occam's Razor."
• The AI paradigm shift changes the game: In the past, logs had relatively low value since they were mostly unstructured text, making them difficult for machines to automatically interpret. But now, AI agents can read all log lines and understand their meanings, extracting unprecedented insights. This enables answering questions that previously required combining metrics and traces solely with logs.

Therefore, Sasabi believes that in the AI era, excellent observability can be achieved by using only logs as telemetry data, greatly simplifying the instrumentation work for engineering teams, allowing them to focus more on building products rather than maintaining complex observability stacks.

This concept is deeply rooted in Shomik's practical experience building observability systems at Brex. As the third infrastructure engineer at Brex, he experienced the company's entire hyper-growth phase from microservices framework, CI/CD systems to production environment setup. As the number of services ballooned to dozens, maintained by different teams, "understanding what happened in production" became a huge challenge, which is precisely the core issue that observability aims to address. He led the team to set up automated instrumentation (automatically adding metrics, tracing, and logs for each microservice), deploying and deeply configuring tools like DataDog, and driving the application of service level goals and metrics. This experience made him acutely aware: "You can do integration tests, unit tests, quality assurance processes, release processes... but nothing you do before the code hits production truly prepares it for production." The production environment is akin to Mike Tyson's famous quote: "Everyone has a plan until they get punched in the mouth." The significance of observability lies in your ability to respond quickly when unpredictable production issues arise.

The Vision of Observability in the AI Era: From Automated Debugging to Self-Repairing Software

Shomik's outlook for Sasabi and the industry is grand and clear. The company's mission is to become the default observability tool for all companies in the “AI era.” He believes that the "AI era" will likely begin around 2025, the "year of the agents," and continue forward.

The longer-term vision is to create a world of "self-repairing software," where software can improve itself without human intervention or guidance. He pointed out that current companies like Cursor are automating the part of software development that involves "creating new features," which is a vast market. However, a larger part of the developer's work—maintaining already written software—is a domain yet to be profoundly transformed by AI, and this is the opportunity that Sasabi aims to capitalize on.

“We absolutely have the opportunity to become a truly incredible, groundbreaking company,” Shomik is confident about this. He sees immense potential in deeply integrating AI into the maintenance segment of the software lifecycle, which may redefine the daily work of developers.

Returning to YC: Acceleration, Culture, and Strategic Distribution

As an alumnus who has already successfully entered the YC network and has relevant experience, Shomik chose to rejoin the latest batch of YC teams, a decision that may seem counterintuitive but was made after careful consideration:

• Business Acceleration: Market opportunities are fleeting, and competitors will emerge. YC, as an accelerator, forces teams to act quickly and push products to market through its set intensive deadlines and rhythms, preventing engineering teams from falling into the trap of "always building."
• Culture Shaping: He hopes to infuse some of YC's values into Sasabi's culture, such as rapid delivery and a focus on market validation rhythms.
• Strategic Distribution and Customer Acquisition: This is a very tactical but crucial point. Every YC company is a software company, and every software company needs observability solutions. Shomik plans to follow the successful paths of excellent YC companies like Brex, Deal, and Rippling to effectively sell products to fellow startups, establishing an early customer base and reputation.
• Making Up for Personal Experience: His first YC experience was completed remotely during COVID, missing out on the immersive in-person experience. He longs to speak on the Demo Day stage and engage deeply with fellow entrepreneurs, rectifying this regret.

Insights for Entrepreneurs and Engineers

Reflecting on the entire journey, Shomik stated that he does not wish to give his past self during the Opkit period any advice that would change the final outcome, as all those "biting the bullet" difficult times—knocking on doors, learning how to start a business in adversity—shaped the resilience of him and his team, and those lessons ultimately benefited Sasabi. If he had to give advice, it would be to emphasize: entrepreneurship is a long game. The lessons learned during Opkit, the relationships built with investors, customers, and engineers, and the connections with fellow YC entrepreneurs will all yield compounded returns over a long time. Treating people sincerely and maintaining good relationships is essential because you may need their support five years from now.

Regarding team building, Sasabi is actively hiring. The ideal candidate in Shomik's mind is: a high degree of autonomy, rapid learning speed, a passion for tools (not limited to development tools), and enjoyment of creating "sharp tools." Individuals with entrepreneurial experience are particularly welcome because they are used to proactively solving problems, engaging in unsexy hard work, and being adaptable—especially important in a rapidly evolving AI landscape where future positions and organizational structures remain unclear. Specifically, the company urgently needs top engineers in database/storage/infrastructure (as Sasabi has built its own log ingestion storage system), full-stack product engineers, and front-end/design engineers who value aesthetics, design, and details. Shomik hopes to mold Sasabi into "the Linear of the observability field," offering an extremely elegant user experience to distinguish it from the complexity of existing tools.

Finally, Shomik hopes his story can inspire other entrepreneurs to find their own "tombstone company"—the one that deeply aligns with their personal identity, passion, and expertise, and motivates them to fully commit to building it. He believes that with all the lessons and advantages gained from past experiences, Sasabi is indeed his "tombstone company."

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。